摘要
近年来,设计高速的网络设备用以在流层面进行分组处理,一直是工业界和学术界的研究热点。10G网络下的TCP流处理同时面临性能问题和异常情况的处理。真实流量的许多特性,会导致基于硬件的TCP流处理系统内存耗尽。针对这种情况,提出并实现了一种用于10G网络中TCP流处理的硬件设计方法,能够处理百万量级TCP流的分组重组和状态跟踪。该解决方案提出了大规模流表的流替换算法和对全局流表的快速的单轮访问方式,采取了对乱序数据缓冲管理的主动释放策略,设计了一种无链的数据结构,保证了最坏情况下的处理时间要求。仿真结果显示,该系统可以处理超过99%的10G网络流量。基于FPGA的硬件原型也验证了系统的设计。
There is an increasing interest in designing high-speed network devices to perform packet processing at flow level.However,TCP processing for 10 Gbps network traffic needs not only to address performance problems but also to cope with abnormal conditions. Some characteristics of real traffic,will result in memory exhaustion in hardware-based TCP processing subsystem which is less flexible for exceptional processing.For this case,this paper presents a hardware design which is capable of processing real traffic in10 G networks with TCP reassembly and tracking states of millions of parallel TCP flows. The solution has several features: 1a simple and effective implementation of flow replacement algorithm for massive flow table; 2 fast one round access to global flow table; 3 an active release policy for out-of-order data buffer management; 4 a link-less data structure which ensures time limit for worst case processing. The simulation result shows that the system can process over 99% of the 10 G network traffic.A FPGA-based prototype is also implemented for evaluation.
出处
《无线电通信技术》
2016年第3期70-75,114,共7页
Radio Communications Technology
基金
国家高技术研究发展计划(863计划)项目(2013AA122105)